Abstract
We propose a compact snapshot monocular depth estimation technique thatrelies on an engineered point spread function (PSF). Traditional approachesused in microscopic super-resolution imaging such as the Double-Helix PSF(DHPSF) are ill-suited for scenes that are more complex than a sparse set ofpoint light sources. We show, using the Cram\'er-Rao lower bound, thatseparating the two lobes of the DHPSF and thereby capturing two separate imagesleads to a dramatic increase in depth accuracy. A special property of the phasemask used for generating the DHPSF is that a separation of the phase mask intotwo halves leads to a spatial separation of the two lobes. We leverage thisproperty to build a compact polarization-based optical setup, where we placetwo orthogonal linear polarizers on each half of the DHPSF phase mask and thencapture the resulting image with a polarization-sensitive camera. Results fromsimulations and a lab prototype demonstrate that our technique achieves up to$50\%$ lower depth error compared to state-of-the-art designs including theDHPSF and the Tetrapod PSF, with little to no loss in spatial resolution.